sorry i forgot to followup here.

El dimarts, 9 de gener del 2024, a les 1:06:57 (CEST), Benjamin Bruce va 
escriure:
> Thanks for your reply. Here are my answers to your questions:
> > Is there an actual standard? Wikipedia tells me there's like several
> > somewhat disagreeing variants.
> 
> Yes, the original Shavian alphabet was the winner of a contest put on by the
> estate of George Bernard Shaw to invent a new phonetic script for English.
> Even though its inventor continued to tinker with it after the contest was
> over, I think it's safe to say that the alphabet as published in Androcles
> and the Lion in 1962
> (https://miro.medium.com/v2/1*ECKikjb6MOfoxV9B8nQHhA.png) is as close to an
> official standard as one can get with such things.

Ok, makes sense.

> > Is the translation possible to be made automatic or needs a human
> > translator?
> Yes, thankfully automatic translation is possible, although it should always
> be human-checked afterwards. There is a Python script I use to convert text
> files to Shavian.

This is a good thing, we would want this to be as automatic as possible.

> > Someone would need to come up with an @script modifier and test if it
> > works, for example for cyrillic we use @cyrillic, e.g. uz@cyrillic and
> > things don't break, one would need to see if setting a glibc-locale like
> > en@shavian would make everything explode or not.
> 
> I don't know if this is useful or not, but Shavian does have a code in ISO
> 15924 (en-Shaw).

Doesn't immediately help, a locale format is
   language[_territory][.codeset][@modifier]
so as far as i can see the thing we can use here is @modifier.

Now, for this to be able to move forward I think you should try to bootstrap 
this with something like a 2 files translation and see if you can get things 
to show up translated.

Do you have knowledge of how the gettext translation system to try to create a 
few files in you system to test this?

Cheers,
  Albert